Two sites, two different claims#
Guests arrive in Georgia having read “8,000 years of wine”, then land in Armenia a week later to hear “6,100 years” and the words “oldest winery in the world”. They ask me which is true. Both are, and they are not competing figures. The Georgian date comes from Neolithic pottery at Gadachrili Gora and Shulaveris Gora in the Marneuli plain of Kvemo Kartli, fifty kilometres south of Tbilisi. The Armenian date comes from the Areni-1 cave in Vayots Dzor, where a complete wine-making installation was dug out of the floor. One site proves that wine existed; the other shows how it was made. I take groups to the second and tell them about the first; this is what I say at the table, with the caveats.
Gadachrili Gora: Neolithic jars south of Tbilisi#
Gadachrili Gora is a low mound in farmland, one of a cluster of early farming villages that archaeologists group as the Shulaveri-Shomu culture. The people lived in round mud-brick houses, grew wheat and barley, kept animals and, from about 6000 BC, fired large clay jars. The mound was first dug in the 1960s and 1970s; a joint Georgian National Museum and University of Toronto team returned in 2012 and, with sensible humour, called the project GRAPE.
In November 2017 the team, with the American archaeological chemist Patrick McGovern, published its results in the Proceedings of the National Academy of Sciences. Sherds from the bases of jars at both Gadachrili Gora and neighbouring Shulaveris Gora carried tartaric acid absorbed into the clay, together with malic, succinic and citric acids. The layers they came from were radiocarbon dated to roughly 6000 to 5800 BC. That pushed the earliest chemical evidence for grape wine back six or seven centuries from the previous holder, Hajji Firuz Tepe in Iran’s Zagros mountains, around 5400 BC.
There is a detail I like more than the headline. One large jar from Khramis Didi Gora, another village in the same group, carries applied clay decoration that has been read as bunches of grapes. Nobody can prove what the potter meant, but a farmer who decorates a storage jar with fruit is telling you what went into it.
What tartaric acid can and cannot say#
Residue chemistry is the heart of both discoveries. Tartaric acid is the signature acid of the grape. In the Near East and the Caucasus few other plants contain it in quantity, so tartaric acid in a sherd points to grapes. On its own that proves grape, not wine; unfermented juice would leave the same mark. The supporting acids help: succinic acid is a by-product of fermentation, and the four together look like what wine leaves behind.
Then common sense does the rest. Crushed grapes in a warm autumn ferment within days whether anyone wants them to or not. A jar of grape juice left in a Neolithic house would have been wine by the following week; the distinction that matters to a chemist barely exists for a farmer.
Nor does the chemistry give a date. Clay cannot be radiocarbon dated; the dates come from charcoal, seeds and bone in the same layer, so the number is the age of the floor the pot stood on. Archaeologists say so in their papers; the marketing copy that follows usually drops it.
Areni-1: a Copper Age winery in a cave#
Drive south from Yerevan for an hour and forty minutes, turn off below Areni village onto the road to Noravank, and the cave opens in the cliff near the mouth of the gorge. From 2007 a team led by Boris Gasparyan, Gregory Areshian and Ron Pinhasi excavated its front chambers. They found a shallow clay basin about a metre across, set with a lip that drained into a jar sunk into the floor beside it. Around it were storage jars, grape seeds, pressed skins, dried vine stems and a drinking cup. The layer dates to around 4100 to 4000 BC, the Late Chalcolithic, the Copper Age.
Chemistry followed in 2011, in the Journal of Archaeological Science. Residues from the basin and the jar contained malvidin, the pigment that makes red grapes red and red wine red. Malvidin turns up in a few other fruits, pomegranate among them, so the chemistry alone would not have closed the case. The seeds and skins on the floor did: they were Vitis vinifera, the wine grape, and there is no other reason to tread grapes in a basin and run the juice into a buried jar.
What makes Areni-1 different is that it is a process, not a trace. You can stand in front of the basin and see the work: grapes trodden by foot, juice running downhill into clay, the jar sealed and left in a cool cave. It is the logic of a karas sunk into a storeroom floor at Karmir Blur two and a half thousand years later, and not far from what a village family in Rind does today.
The cave also held graves, and the winery sits among them; the excavators suggested the wine may have been made for burial rites rather than for dinner. The same chambers gave up the oldest known leather shoe, about 5,500 years old, a moccasin stuffed with grass. On tour the shoe gets more questions than the vat does.
What the pots cannot tell us#
Neither site tells us who made wine first. Anything fermented in a wooden trough, a skin or a gourd before the first jars were fired left nothing for a chemist to find. The pots mark the moment the evidence starts, not the moment the practice did.
The vines say something similar. A 2023 study in Science sequenced the genomes of several thousand cultivated and wild grapevines and argued that the vine was domesticated twice, around eleven thousand years ago, once in the South Caucasus and once in Western Asia. If that holds, people were selecting grapes three thousand years before the Gadachrili jars. The eight thousand years on the Georgian posters is a floor, not a birthday.
Neither site settles the argument my guests most want settled. The two places are about two hundred kilometres apart, across a border that did not exist, in a region where the wild vine still grows in the river woods. I speak Armenian, I take groups to both countries, and I decline to referee: the South Caucasus, taken as a whole, is where the record begins.
From the Neolithic jar to the qvevri and the karas#
The line from those jars to a working cellar is clearest in Georgia. The qvevri, egg-shaped, beeswax-lined and buried to the neck, is the descendant of the Shulaveri jars in shape and function. Eleventh-century qvevri are still set in the floor of the marani at Alaverdi Monastery in Kakheti, and monks make wine there again. Village families kept qvevri through every kingdom and the Soviet decades, when the state made wine in tanks and the yard made it in clay. In 2013 UNESCO listed the qvevri method as intangible cultural heritage, which is the bureaucratic way of saying the practice never stopped. The nearest working qvevri cellar I visit to the Neolithic mounds is Iago Bitarishvili’s in Chardakhi, in Kartli north-west of Tbilisi, where old qvevri hold one grape, Chinuri.
Armenia’s line broke and is being mended. The Urartian fortress of Karmir Blur, inside modern Yerevan, kept wine in the seventh century BC in storerooms of karas sunk to the shoulder in the floor, many marked with their capacity. Then Soviet planning made Armenia the brandy republic, the Ararat plain went to brandy grapes, and the karas left the winery for the village yard. The estates founded after 2006 went and found them there. Zorik Gharibian at Zorah bought karas from village families and buried them at his winery in Rind, twenty minutes from the cave; Trinity Canyon Vineyards in Aghavnadzor named its main range 6100 after the age of the Areni-1 winery and ferments its Ancestors line in karas. The vessels are old, the practice is being relearned, and both countries are doing the same thing: reading their own pots.
On tour#
On my Armenia tour we walk into Areni-1 on the way to Noravank. It is a short walk from the road, there is a small fee, and the excavated area is near the entrance. I bring a torch and a bottle of unoaked Areni Noir from a cellar a few kilometres away, and we drink it outside with the gorge in front of us. Nobody alive has tasted a Chalcolithic wine and I make no claim about what it was like. But the grape is the same species and the valley is the same valley.
Gadachrili Gora I describe rather than visit: a mound in a field near Marneuli with nothing for a group to see. The sherds that carried the tartaric acid are in the care of the Georgian National Museum in Tbilisi, and on the two cradles route we go from the Areni cave to that museum in the same week, then drink an amber Rkatsiteli from a buried qvevri in Kakheti that evening. Put the three in a row, the Neolithic sherd, the Copper Age basin and tonight’s glass, and the eight thousand years stop being a slogan. They become a working method that village families never put down, and that a new generation in Vayots Dzor is picking back up. For how a qvevri differs from a karas, see qvevri, karas and amphora.
